Optimization of Material Composition of Li-Intercalated Metal-Organic Framework Electrodes Using a Combination of Experiments and Machine Learning of X-Ray Diffraction Patterns

Although individual Li-intercalated metal-organic frameworks (iMOFs) exhibit properties that make them suitable for use as electrodes in Li-ion batteries, none of these materials have the ideal combination of electrochemical characteristics. Using the high specific capacity and low polarization of 2,6-naphthalene dicarboxylate dilithium as a basis for improvement, 26 iMOFs from different combinations of aqueous terephthalic, 2,6-naphthalene dicarboxylic, and 4,4 '-biphenyl dicarboxylic-based solutions with Li source are synthesized to obtain a material with the optimal electrochemical performances. For a more comprehensive search of the optimal ratio of raw material solutions, a machine learning-based prediction model is constructed, using a combination of X-ray diffraction (XRD) patterns and the experimentally derived characteristics of the 26 iMOFs. With this model, the optimal iMOF ratio is found to be a 2,6-naphthalene dicarboxylic acid-rich ternary solution of raw material. It is concluded that the improvement in the electrochemical properties originates from a change in the iMOFs crystal structure caused by synthesis from three solutions. As the model is independent of fabrication parameters such as heating temperature, it can also be used for evaluation of synthesis procedures. Hence, the XRD data-based machine learning method introduced in this study is a powerful tool for practical material development.